Sewage pipe network flow prediction algorithm based on rainfall and liquidometer data

Abstract

The invention discloses a sewage pipe network flow prediction algorithm based on rainfall and liquidometer data, and belongs to the field of flow prediction.The sewage pipe network flow prediction algorithm based on rainfall and liquidometer data comprises the following steps that historical and real-time multi-dimensional data from multiple sources are continuously collected and integrated; performing cleaning, calibration and time alignment processing on the multi-dimensional data to obtain model training data based on historical data and model prediction input data based on real-time data; the method has the advantages that multi-dimensional data are collected and processed systematically, fuzzy manual observation is replaced, historical hydrological laws are converted into a quantifiable sewage pipe network flow prediction model, judgment depending on personal experience is replaced, accurate flow prediction in several hours in the future is carried out based on the model, and the flow prediction efficiency is improved. Passive emergency is converted into active early warning, and the hysteresis and subjective defects caused by dependence on artificial experience in the prior art are overcome fundamentally.

Description

Technical Field

[0001] This invention belongs to the field of flow prediction, and in particular relates to a flow prediction algorithm for sewage pipe networks based on rainfall and level gauge data. Background Technology

[0002] Currently, during rainfall, the operation and management of sewage pipe networks mainly rely on the on-site experience and judgment of the personnel. Staff decide when to activate or deactivate critical facilities such as storage tanks and pumping stations based on the observed rainfall intensity. This model, dependent on manual decision-making, suffers from significant lag and subjectivity.

[0003] Due to the lack of accurate prediction of flow changes within the pipeline network, the timing for initiating response measures is often difficult to grasp. Delayed decision-making or misjudgment can easily lead to untimely or inappropriate implementation of measures. The direct consequence is an inability to effectively cope with sudden increases in flow, frequently causing sewage overflows that pollute the environment and even exacerbate the risk of urban flooding, posing a threat to public safety and the ecological environment. Improvements are needed. Summary of the Invention

[0004] Therefore, it is necessary to provide a sewage pipe network flow prediction algorithm based on rainfall and level gauge data to address the above problems.

[0005] The present invention is implemented as follows: a sewage pipe network flow prediction algorithm based on rainfall and level gauge data includes the following steps:

[0006] Continuously collect and integrate historical and real-time multi-dimensional data from multiple sources. The multi-dimensional data includes historical and real-time rainfall data (obtained through rain gauges), historical and real-time liquid level readings recorded by liquid level gauges at monitoring nodes in the sewage pipe network, and weather forecast information from meteorological departments (especially predictions of future rainfall intensity and duration). The multi-dimensional data is cleaned, calibrated, and time-aligned to obtain model training data based on historical data and model prediction input data based on real-time data.

[0007] Construct a sewage pipe network flow prediction model (based on machine learning or deep learning algorithms), and use model training data to train the sewage pipe network flow prediction model. The training objective is to enable the sewage pipe network flow prediction model to learn the laws and lag effects of liquid level and flow changes at each monitoring node in the sewage pipe network under different rainfall patterns (such as rainfall intensity, duration, and spatial distribution).

[0008] When rainfall occurs, the model prediction input data (real-time rainfall data, current sewage network liquid level data, and weather forecast data about future rainfall) are input into the pre-trained sewage network flow prediction model. The sewage network flow prediction model outputs the flow and liquid level change prediction curves of each monitoring node in the sewage network within a future period (e.g., the next 1-3 hours).

[0009] In one embodiment, the present invention provides a sewage network flow prediction algorithm based on rainfall and level gauge data, further comprising:

[0010] When it is predicted that the flow rate or liquid level of a certain sewage pipe network monitoring node will reach or exceed a preset first threshold in the future (e.g., in the next 1-3 hours), an early warning mechanism is activated. The mechanism combines the current forecast information of the monitoring node with the subsequent weather forecast rainfall and issues different levels of early warning information (e.g., blue alert, yellow alert, red high-risk alert) through the control center's large screen, relevant personnel's computers and mobile terminals, etc.

[0011] In one embodiment, the present invention provides a sewage network flow prediction algorithm based on rainfall and level gauge data, further comprising:

[0012] When issuing early warning information, based on the same forecast results, subsequent weather forecast rainfall, and the built-in expert knowledge base and preset scheduling rules, the system automatically analyzes and generates targeted response measures and suggestions.

[0013] In one embodiment, the present invention provides a sewage network flow prediction algorithm based on rainfall and level gauge data, further comprising:

[0014] Establish a regional importance database to provide a hierarchical basis for data collection strategies. When issuing early warning information, dynamically adjust the collection frequency of real-time multi-dimensional data according to the hierarchical classification of the regions involved (for example, adjust the collection frequency from once every 5 minutes to once every 1 minute for high-importance region A, and adjust it to once every 3 minutes for region B).

[0015] In one embodiment, the present invention provides a sewage network flow prediction algorithm based on rainfall and level gauge data, further comprising:

[0016] Continuously monitor real-time multi-dimensional data. When a certain data is abnormal, it is marked as abnormal data and removed. Based on the historical data of the monitoring node where the abnormal data is located and the data of the surrounding normal monitoring nodes, the data is filled in by interpolation.

[0017] In one embodiment, the present invention provides a sewage pipe network flow prediction algorithm based on rainfall and level gauge data. The criteria for judging abnormal data are: the difference between the rate of change of the data and the rate of change of the data of surrounding monitoring nodes exceeds a second threshold, or the absolute value of the difference between the data and the data of the same period in history under the same environment exceeds a third threshold.

[0018] In one embodiment, the present invention provides a sewage network flow prediction algorithm based on rainfall and level gauge data, further comprising:

[0019] Integrate comprehensive geographic information and design data of urban drainage pipe networks (sewage pipe networks and stormwater pipe networks) to construct a digital archive containing pipe network types and topology. Based on the digital archive, mark all combined sewer pipe network sections and combined sewer overflow locations within the service area.

[0020] If a combined sewer overflow is detected, a flood discharge prediction module is added to the sewage network flow prediction model. The module is trained using historical flood discharge event data (such as flood discharge intensity, duration, spatial distribution from upstream, etc.) to learn the dramatic changes in flow and liquid level in the combined sewer network under flood discharge conditions.

[0021] When a flood discharge actually occurs, the real-time flood discharge characteristic data (such as flood discharge flow and expected duration) will be input into the flood discharge prediction function module of the sewage pipe network flow prediction model. The flood discharge prediction function module will predict the flow and liquid level change curves of the pipe network nodes in the area directly affected by the flood discharge in the future (e.g., the next 1-3 hours).

[0022] In one embodiment, the present invention provides a sewage network flow prediction algorithm based on rainfall and level gauge data, further comprising:

[0023] Establish a pipeline health status database based on historical dredging records, video inspection reports, and hydraulic performance analysis. Assign a dynamically updated siltation coefficient to each monitoring node pipe segment. When performing flow prediction calculations, automatically correct the Manning roughness coefficient and effective flow area of ​​the corresponding pipe segment based on the siltation coefficient.

[0024] In one embodiment, the present invention provides a sewage network flow prediction algorithm based on rainfall and level gauge data, further comprising:

[0025] By accessing real-time meteorological data and monitoring the number of consecutive drought days, when the sewage pipe network flow prediction model is running, a dynamic adjustment factor is calculated based on the current real-time meteorological data and short-term forecast meteorological data. The sedimentation coefficient of the upstream sewage pipe network (especially the combined sewer pipe network section) is temporarily increased to simulate the impact of the temporary reduction in flow capacity during the initial flushing of the first rain after a long drought. The adjustment factor is a temporary parameter and is only used for a single prediction calculation.

[0026] In one embodiment, the present invention provides a sewage network flow prediction algorithm based on rainfall and level gauge data, further comprising:

[0027] Using the effective flow area and structural score data of the pipeline in the pipeline health status database, a unique critical hydraulic condition is calculated for each monitoring node. Based on the latest measured geometric properties of the pipe section where the monitoring node is located (such as the flow area and slope after correction by the sedimentation coefficient) and material condition, the theoretical liquid level value for the pipe section to reach full flow or pressurized flow state is derived through hydraulic calculation formula. On this basis, combined with the structural defect score of the pipeline, a quantitative safety margin is allocated to the pipeline, thereby generating a personalized fourth threshold.

[0028] Compared with existing technologies, the beneficial effects of this invention are as follows: This invention replaces vague manual observation by systematically collecting and processing multi-dimensional data, transforming historical hydrological patterns into quantifiable sewage pipe network flow prediction models, replacing judgments based on personal experience, and making accurate flow predictions for the next few hours based on the model, realizing the transformation from passive emergency response to proactive early warning, and fundamentally overcoming the lag and subjectivity defects caused by existing technologies relying on human experience. Attached Figure Description

[0029] Figure 1 This is a schematic diagram of the first part of a sewage pipe network flow prediction algorithm based on rainfall and level gauge data, provided as an embodiment of the present invention.

[0030] Figure 2 This is a schematic diagram of the second part of a sewage pipe network flow prediction algorithm based on rainfall and level gauge data, provided as an embodiment of the present invention.

[0031] Figure 3 This is a schematic diagram of the third part of a sewage pipe network flow prediction algorithm based on rainfall and level gauge data, provided as an embodiment of the present invention.

[0032] Figure 4 This is a schematic diagram of the fourth part of a sewage pipe network flow prediction algorithm based on rainfall and level gauge data, provided as an embodiment of the present invention.

[0033] Figure 5 This is a schematic diagram of the fifth part of a sewage pipe network flow prediction algorithm based on rainfall and level gauge data, provided as an embodiment of the present invention.

[0034] Figure 6 This is a schematic diagram of the sixth part of a sewage pipe network flow prediction algorithm based on rainfall and level gauge data, provided as an embodiment of the present invention.

[0035] Figure 7 This is a schematic diagram of the seventh part of a sewage pipe network flow prediction algorithm based on rainfall and level gauge data, provided as an embodiment of the present invention.

[0036] Figure 8 This is a schematic diagram of the eighth part of a sewage pipe network flow prediction algorithm based on rainfall and level gauge data, provided as an embodiment of the present invention.

[0037] Figure 9 This is a schematic diagram of the ninth part of a sewage pipe network flow prediction algorithm based on rainfall and level gauge data, provided as an embodiment of the present invention. Detailed Implementation

[0038] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0039] It is understood that the terms "first," "second," etc., used in this application may be used herein to describe various elements, but unless otherwise specified, these elements are not limited by these terms. These terms are used only to distinguish one element from another. For example, without departing from the scope of this application, a first script may be referred to as a second script, and similarly, a second script may be referred to as a first script.

[0040] In one embodiment, such as Figure 1 As shown, a sewage pipe network flow prediction algorithm based on rainfall and level gauge data includes the following steps:

[0041] Step S1: Continuously collect and integrate historical and real-time multi-dimensional data from multiple sources. The multi-dimensional data includes historical and real-time rainfall data (obtained through rain gauges), historical and real-time liquid level readings recorded by liquid level gauges at monitoring nodes in the sewage pipe network, and weather forecast information from meteorological departments (especially predictions of future rainfall intensity and duration). The multi-dimensional data is cleaned, calibrated, and time-aligned to obtain model training data based on historical data and model prediction input data based on real-time data.

[0042] Step S2: Construct a sewage pipe network flow prediction model (based on machine learning or deep learning algorithms), and train the sewage pipe network flow prediction model using model training data. The training objective is to enable the sewage pipe network flow prediction model to learn the patterns and lag effects of changes in liquid level and flow at each monitoring node in the sewage pipe network under different rainfall patterns (such as rainfall intensity, duration, and spatial distribution).

[0043] Step S3: When rainfall occurs, input the model prediction data (real-time rainfall data, current sewage network liquid level data, and weather forecast data about future rainfall) into the trained sewage network flow prediction model. The sewage network flow prediction model outputs the flow and liquid level change prediction curves of each monitoring node of the sewage network within a future period (e.g., the next 1-3 hours).

[0044] The construction of a wastewater pipe network flow prediction model is illustrated using a deep learning framework, specifically employing a Long Short-Term Memory (LSTM) network, a model suitable for time series analysis. First, feature engineering is performed on the data processed in step S1, extracting rainfall intensity, duration, spatial distribution, and pipe network topology as input features, while historical liquid level and flow sequences are used as supervision labels. The model structure includes an input layer, multiple LSTM hidden layers, and an output layer. The LSTM layers learn the nonlinear relationship and time lag effect between rainfall events and the pipe network hydraulic response through a gating mechanism. During training, gradient descent is used to optimize the loss function, and backpropagation is used to continuously adjust the network parameters, enabling the model to accurately predict future trends in liquid level and flow at each monitoring point.

[0045] Steps S1-S3 constitute a closed-loop iterative optimization process. During the initial system construction and operation phase, the historical data processed in step S1 is used as model training data for the initial model training in step S2. When a rainfall event occurs, the real-time data processed in step S1 is used as model prediction input data for real-time prediction in step S3. After the prediction is completed, the complete model prediction input data for that event and its corresponding actual monitoring results (such as actual flow rate and liquid level) will be automatically collected, aligned, and labeled by the system to form new, timely training samples. These new samples will be incorporated into the historical database and used as model training data in the next training cycle, thereby achieving continuous updating and optimization of the sewage pipe network flow prediction model, enabling it to dynamically adapt to changes in the pipe network and improve prediction accuracy.

[0046] In one embodiment, such as Figure 2 As shown, a sewage pipe network flow prediction algorithm based on rainfall and level gauge data also includes:

[0047] Step S4: When it is predicted that the flow rate or liquid level of a certain sewage pipe network monitoring node will reach or exceed the preset first threshold in the future (e.g., in the next 1-3 hours), the early warning mechanism is activated. Based on the current forecast information of the monitoring node and the subsequent weather forecast rainfall, different levels of early warning information (e.g., blue alert, yellow alert, red high-risk alert) are issued (through the large screen of the control center, the computers and mobile terminals of relevant personnel, etc.).

[0048] The first threshold is mainly determined based on the theoretical full-flow capacity of the pipeline under design conditions. For example, by using hydraulic calculation formulas, combined with the design pipe diameter, slope and Manning roughness coefficient, the theoretical flow rate and liquid level value when it reaches full flow (i.e., the fullness is close to 1.0) are calculated, and a safety margin (such as 90%) is set on this theoretical value as the first threshold.

[0049] In one embodiment, such as Figure 3 As shown, a sewage pipe network flow prediction algorithm based on rainfall and level gauge data also includes:

[0050] Step S5: When issuing early warning information, based on the same forecast results, subsequent weather forecast rainfall, and the built-in expert knowledge base and preset scheduling rules, automatically analyze and generate targeted response measures suggestion text.

[0051] For example, it might explicitly state, "It is recommended to activate pump station XX in 20 minutes to reduce downstream pressure" or "It is recommended to immediately close gate XX to guide water flow to the storage tank." These specific and actionable instructions help managers make more timely and reasonable decisions, proactively mitigating the risks of overflows and flooding.

[0052] In one embodiment, such as Figure 4 As shown, a sewage pipe network flow prediction algorithm based on rainfall and level gauge data also includes:

[0053] Step S6: Establish a regional importance database to provide a hierarchical basis for data collection strategies. When issuing early warning information, dynamically adjust the collection frequency of real-time multi-dimensional data according to the hierarchical classification of the regions involved (for example, adjust the collection frequency from once every 5 minutes to once every 1 minute for high-importance region A, and adjust it to once every 3 minutes for region B).

[0054] Area A (High Importance Area) includes, for example: the direct upstream catchment area containing core facilities such as sewage treatment plants, important pumping stations, and regulating reservoirs; key areas that have experienced multiple sewage overflows or flooding incidents in the past with a wide impact; critical bottlenecks in the pipeline network topology or downstream main pipelines; and corresponding underground pipeline sections with sensitive targets on the surface, such as hospitals, schools, transportation hubs, and water source protection areas.

[0055] Area B (Generally Important Areas): For example, branch pipe networks in residential areas with few historical overflow records; ordinary areas that are mainly diverted and whose ends merge into the main pipe.

[0056] In one embodiment, such as Figure 5 As shown, a sewage pipe network flow prediction algorithm based on rainfall and level gauge data also includes:

[0057] Step S7: Continuously monitor real-time multi-dimensional data. When a certain data is abnormal, mark the data as abnormal data and remove the abnormal data. Based on the historical data of the monitoring node where the abnormal data is located and the data of the surrounding normal monitoring nodes, fill the data by interpolation.

[0058] Sensors that are clogged with leaves or debris, or whose readings drift, may provide incorrect data (e.g., the level gauge reading remains unchanged for a long time or fluctuates violently).

[0059] For example, data completion through interpolation can be implemented when a level gauge is blocked. By combining historical level-flow relationships at that point with readings from upstream and downstream normal level gauges, a machine learning model (such as a regression algorithm) can be used to calculate the most probable level value at that point in real time. In this way, even during short-term sensor malfunctions, continuous and reasonable data input is provided, maximizing the continuity of flow prediction and buying valuable time for decision-making until the faulty sensor is repaired.

[0060] In one embodiment, a sewage pipe network flow prediction algorithm based on rainfall and level gauge data is used to determine abnormal data as follows: the difference between the rate of change of the data and the rate of change of the data of surrounding monitoring nodes exceeds a second threshold, or the absolute value of the difference between the data and the data of the same period in history under the same environment exceeds a third threshold.

[0061] The second threshold (for differences in data change rates) is calculated by analyzing the distribution of the difference in the rate of change of liquid level or flow between the target monitoring node and its neighboring monitoring nodes during the same time period in historical normal events (such as rainfall) (e.g., calculating its standard deviation σ). The threshold is usually set as a specific multiple of this distribution range (e.g., ±3σ). If the difference between the instantaneous rate of change of a node and the mean rate of change of the surrounding nodes exceeds this range, it is judged as abnormal.

[0062] The third threshold (for absolute value differences in data) is formed by analyzing the liquid level or flow rate readings of the target monitoring node in historical events with similar rainfall intensity and duration to the current situation, to create an expected value range (such as the interval formed by the 5th percentile to the 95th percentile). If the current real-time data deviates from this historical expected value range by more than a preset tolerance ratio (such as exceeding the upper limit of the range by 20%), it is judged as abnormal.

[0063] The initial values ​​for both thresholds are derived from historical data statistics, and can be further calibrated and optimized based on false alarms and missed alarms after the system is running.

[0064] In one embodiment, such as Figure 6 As shown, a sewage pipe network flow prediction algorithm based on rainfall and level gauge data also includes:

[0065] Step S8: Integrate comprehensive geographic information and design data of urban drainage pipe network (sewage pipe network and rainwater pipe network) to construct a digital archive containing pipe network type and topology, and mark all combined sewer pipe network sections and combined sewer overflow locations within the service area based on the digital archive;

[0066] Step S9: If a combined sewer overflow outlet is identified, a flood discharge prediction function module is added to the sewage pipe network flow prediction model. The module is trained using historical flood discharge event data (such as flood discharge intensity, duration, spatial distribution from upstream, etc.) to learn the dramatic change patterns of flow and liquid level in the combined sewer network under flood discharge conditions.

[0067] Step S10: When a flood discharge actually occurs, the real-time acquired flood discharge characteristic data (such as flood discharge flow rate and estimated duration) is input into the flood discharge prediction function module of the sewage pipe network flow prediction model. The flood discharge prediction function module then predicts the flow rate and liquid level change curves of the pipe network nodes in the area directly affected by the flood discharge over a future period (e.g., the next 1-3 hours). This provides managers with specific and accurate prediction information for this extreme risk, compensating for the prediction blind spots of single rainfall models in such scenarios.

[0068] In many old urban areas, a combined sewer system is in place, meaning rainwater and sewage share the same pipe network. Normally, sewage is sent to treatment plants; during rainfall, if the volume of sewage mixed with rainwater exceeds a certain limit, it is discharged directly into natural water bodies through a dedicated combined sewer overflow to prevent overpressure on the pipe network. In extreme situations like flood discharge, this overflow will continuously discharge large volumes, but the upstream combined sewer system itself is already filled with floodwater far exceeding its design capacity. The immense water pressure may cause floodwater to seep into the ground through weak points such as basements and pipe gallery joints, eventually finding its way into the sewage network. Therefore, it is necessary to predict the sewage network flow rate under the specific circumstances of flood discharge.

[0069] In one embodiment, such as Figure 7 As shown, a sewage pipe network flow prediction algorithm based on rainfall and level gauge data also includes:

[0070] Step S11: Establish a pipeline health status database based on historical dredging records, video inspection reports, and hydraulic performance analysis. Assign a dynamically updated siltation coefficient to each monitoring node pipe segment. When performing flow prediction calculations, automatically correct the Manning roughness coefficient and effective flow area of ​​the corresponding pipe segment based on the siltation coefficient.

[0071] Furthermore, for severely silted pipe sections, a higher surface roughness and a smaller calculated pipe diameter can be used. This accurately reflects the decrease in water flow capacity caused by sediment accumulation in the pipeline, effectively avoiding prediction deviations caused by the discrepancy between theoretical and actual capacity in silted pipe networks, and significantly improving the reliability of prediction results under non-ideal operating conditions.

[0072] In one embodiment, such as Figure 8 As shown, a sewage pipe network flow prediction algorithm based on rainfall and level gauge data also includes:

[0073] Step S12: Access real-time meteorological data and monitor the number of consecutive drought days. When the sewage pipe network flow prediction model is running, calculate a dynamic adjustment factor based on the current real-time meteorological data and short-term forecast meteorological data. Temporarily increase the sedimentation coefficient of the upstream sewage pipe network (especially the combined sewer pipe network section) to simulate the impact of the temporary reduction in flow capacity during the initial flushing of the first rain after a long drought. The adjustment factor is a temporary parameter and is only used for a single prediction calculation.

[0074] The rate of sediment accumulation in pipelines is not constant. During the dry season, the long-term low flow velocity makes it easy for sediment to settle; after a long drought, the first rain will wash up a large amount of the previous sediment, forming a "first flushing" effect.

[0075] For example, when predicting the first rainfall event after a prolonged drought, this adjustment factor is applied to the baseline sedimentation coefficient of the relevant pipe section to simulate the temporary decrease in flow capacity caused by the "first flush" effect. This adjustment factor is a temporary parameter used only for a single prediction calculation and is reset after the simulation is completed, thereby achieving a more refined characterization of the instantaneous hydraulic state of the pipe network.

[0076] In one embodiment, such as Figure 9 As shown, a sewage pipe network flow prediction algorithm based on rainfall and level gauge data also includes:

[0077] Step S13: Using the effective flow area and structural score data of the pipeline in the pipeline health status database, calculate the exclusive critical hydraulic conditions for each monitoring node. Based on the latest measured geometric properties of the pipe section where the monitoring node is located (such as the flow area and slope after correction by the sedimentation coefficient) and material condition, derive the theoretical liquid level value for the pipe section to reach full flow or pressurized flow state through hydraulic calculation formula. On this basis, combined with the structural defect score of the pipeline, allocate a quantitative safety margin for the pipeline, thereby generating a personalized fourth threshold.

[0078] A uniform liquid level warning threshold (e.g., 2.5 meters) is unscientific for pipelines in different health states. A healthy new pipe may still be safe at 2.5 meters, while an old pipe with a siltation rate of 40% may be close to full flow or even overflow at 2.2 meters. The fourth threshold will be directly used for warning judgment and completely replace the first threshold in step S4. This changes the warning triggering standard from a static value based on idealized design to a dynamic value based on the individualized actual health state of the pipeline, thereby significantly improving the accuracy and reliability of the warning.

[0079] It should be understood that although the steps in the flowcharts of the various embodiments of the present invention are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the various embodiments may include multiple sub-steps or multiple stages. These sub-steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these sub-steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least a portion of the sub-steps or stages of other steps.

[0080] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0081] The embodiments described above are merely illustrative of several implementations of the present invention, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of the present invention. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of the present invention, and these modifications and improvements all fall within the scope of protection of the present invention. Therefore, the scope of protection of this patent should be determined by the appended claims.

[0082] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

[0083] Furthermore, it should be understood that although this specification describes embodiments, not every embodiment contains only one independent technical solution. This narrative style is merely for clarity. Those skilled in the art should consider the specification as a whole, and the technical solutions in each embodiment can also be appropriately combined to form other embodiments that can be understood by those skilled in the art.

Claims

1. A sewage pipe network flow prediction algorithm based on rainfall and level gauge data, characterized in that, The wastewater network flow prediction algorithm based on rainfall and level gauge data includes the following steps: Continuously collect and integrate historical and real-time multi-dimensional data from multiple sources, including historical and real-time rainfall data, historical and real-time liquid level readings recorded by level gauges at monitoring nodes in the sewage pipe network, and weather forecast information from meteorological departments. Clean, calibrate, and time-align the multi-dimensional data to obtain model training data based on historical data and model prediction input data based on real-time data. A sewage pipe network flow prediction model is constructed, and the model training data is used to train the sewage pipe network flow prediction model. The training objective is to enable the sewage pipe network flow prediction model to learn the laws and lag effects of the changes in liquid level and flow at each monitoring node in the sewage pipe network under different rainfall patterns. When rainfall occurs, the model prediction input data is fed into the pre-trained sewage network flow prediction model, which then outputs the predicted flow and liquid level changes of each monitoring node in the sewage network over a future period.

2. The sewage network flow prediction algorithm based on rainfall and level gauge data according to claim 1, characterized in that, Also includes: When it is predicted that the flow rate or liquid level of a certain sewage pipe network monitoring node will reach or exceed a preset first threshold in the future, an early warning mechanism is activated. Based on the current forecast information of the monitoring node and the subsequent weather forecast rainfall, different levels of early warning information are issued.

3. The sewage pipe network flow prediction algorithm based on rainfall and level gauge data according to claim 2, characterized in that, Also includes: When issuing early warning information, based on the same forecast results, subsequent weather forecast rainfall, and the built-in expert knowledge base and preset scheduling rules, the system automatically analyzes and generates targeted response measures and suggestions.

4. The sewage network flow prediction algorithm based on rainfall and level gauge data according to claim 2 or 3, characterized in that, Also includes: Establish a regional importance database to provide a hierarchical basis for data collection strategies. When issuing early warning information, dynamically adjust the collection frequency of real-time multi-dimensional data according to the hierarchical classification of the regions involved.

5. The sewage pipe network flow prediction algorithm based on rainfall and level gauge data according to claim 1, characterized in that, Also includes: Continuously monitor real-time multi-dimensional data. When a certain data is abnormal, it is marked as abnormal data and removed. Based on the historical data of the monitoring node where the abnormal data is located and the data of the surrounding normal monitoring nodes, the data is filled in by interpolation.

6. The sewage network flow prediction algorithm based on rainfall and level gauge data according to claim 5, characterized in that, The criteria for judging abnormal data are: the difference between the rate of change of the data and the rate of change of the data of surrounding monitoring nodes exceeds the second threshold, or the absolute value of the difference between the data and the data of the same period in history under the same environment exceeds the third threshold.

7. The sewage network flow prediction algorithm based on rainfall and level gauge data according to claim 1, characterized in that, Also includes: Integrate comprehensive geographic information and design data of urban drainage pipe networks to construct a digital archive containing pipe network types and topology. Based on the digital archive, mark all combined sewer pipe network sections and combined sewer overflow locations within the service area. If a combined sewer overflow is detected, a flood discharge prediction module is added to the sewage network flow prediction model. This module is trained using historical flood discharge event data to learn the dramatic changes in flow and liquid level within the combined sewer network under flood discharge conditions. When a flood discharge actually occurs, the real-time flood discharge characteristic data is input into the flood discharge prediction function module of the sewage pipe network flow prediction model. The flood discharge prediction function module predicts the flow and liquid level change curves of the pipe network nodes in the area directly affected by the flood discharge in the future.

8. The sewage network flow prediction algorithm based on rainfall and level gauge data according to claim 1 or 7, characterized in that, Also includes: Establish a pipeline health status database based on historical dredging records, video inspection reports, and hydraulic performance analysis. Assign a dynamically updated siltation coefficient to each monitoring node pipe segment. When performing flow prediction calculations, automatically correct the Manning roughness coefficient and effective flow area of ​​the corresponding pipe segment based on the siltation coefficient.

9. The sewage network flow prediction algorithm based on rainfall and level gauge data according to claim 8, characterized in that, Also includes: By accessing real-time meteorological data and monitoring the number of consecutive drought days, when the sewage pipe network flow prediction model is running, a dynamic adjustment factor is calculated based on the current real-time meteorological data and short-term forecast meteorological data to temporarily increase the siltation coefficient of the upstream sewage pipe network, simulating the impact of the temporary reduction in flow capacity during the initial flushing of the first rain after a long drought. The adjustment factor is a temporary parameter and is only used for a single prediction calculation.

10. The sewage pipe network flow prediction algorithm based on rainfall and level gauge data according to claim 8, characterized in that, Also includes: Using the effective flow area and structural score data of the pipeline in the pipeline health status database, a unique critical hydraulic condition is calculated for each monitoring node. Based on the latest measured geometric properties and material conditions of the pipe section where the monitoring node is located, the theoretical liquid level value for the pipe section to reach full flow or pressurized flow state is derived through hydraulic calculation formula. On this basis, combined with the structural defect score of the pipeline, a quantitative safety margin is allocated to the pipeline, thereby generating a personalized fourth threshold.

Images